Off the shelf, onto the beach: commercial GPS in amphibious combat vehicles

GPS World, April, 2004 by Stephanie Edwards, Marlin Gendron, Maura Lohrenz, Richard Mang

Although the PLGR was used as the baseline for testing, it is not always available to every AAV crew in either training or wartime environments. Moreover, the crew members in the NRL trials exhibited unfamiliarity with its function, which required additional time to train them in PLGR operation. After the initial PLGR training, the NRL team spent about ten minutes explaining the moving map concept and instructing drivers on its operation.

Each test or demonstration took place on a predetermined course based on the area in which the vehicles were cleared to operate. Specific waypoints were entered into both the moving map system and the PLGR. The PLGR showed position numerically, while the moving map system showed position graphically.

When navigating with the moving-map display, AAV drivers were instructed to follow the lane markings on the display and to stay as close to the centerline as possible. When navigating with the PLGR, AAV drivers were told to aim for the next waypoint as precisely as possible. The moving-map display was turned off during PLGR tests, and PLGR were not issued to drivers during moving map tests. Both test conditions (moving map and PLGR) were repeated with the same drivers on the same course, in both clockwise and counterclockwise directions to reduce familiarity. These runs were repeated over several days, with vehicle positions recorded once per second by the NRL moving map system's computer for later analysis.

[FIGURE 3 OMITTED]

Results

Test results calculated how well the drivers could stay in their lanes using the moving map compared to results when using the PLGR. This was accomplished by comparing each individual run to the actual course measured in terms of cross track error (CTE), which is the positive perpendicular distance between the planned route and the actual track (recorded as a series of latitude and longitude points from the DGPS receiver), and is similar in magnitude to root mean square error:

CT[E.sub.P] = [absolute value of [[C.sub.X][C.sub.Y]] * [([Y.sub.E]-[Y.sub.S])([X.sub.P]-[X.sub.S]) - ([X.sub.E]-[X.sub.S])([Y.sub.P]-[Y.sub.S])]]/[SQRT [([C.sub.X] ([X.sub.E]-[X.sub.S])[).sup.2] ([C.sub.Y]([Y.sub.E]-[Y.sub.S])[).sup.2]]]

Where:

[C.sub.X] = constant to convert longitude into meters (for the average latitude of the course),

[C.sub.Y] = constant to convert latitude into meters (which is independent of longitude),

([X.sub.P], [Y.sub.P]) = longitude (X) and latitude (Y) of the DGPS point along the actual track,

([X.sub.S], [Y.sub.S]) = longitude and latitude of the starting point of the planned route segment, and

([X.sub.E], [Y.sub.E]) = longitude and latitude of the ending point of the planned route segment.

The CTE for the entire track is calculated as the average of the individual CTEs for all points recorded along the track, which is broken into turns and straight sections. For better comparisons, average CTE values are calculated separately for each type of section.

[FIGURE 4 OMITTED]

The drivers who had experience using a PLGR were reluctant to accept that the moving-map display might improve their lane navigation performance. However, even the experienced driver of the track shown in Figure 3 experienced a common PLGR problem: missing a waypoint. When a waypoint is accidentally missed while using a PLGR, the driver can only aim for the next waypoint. Using PLGR navigation, a driver has no way to regain the track until the AAV reaches the next waypoint. This creates a potentially dangerous situation, because the AAV runs the risk of hitting a mine whenever it is outside the predetermined lane. The longer it remains outside the lane, the more risk it faces.